Improving Human Activity Classification Based on Micro-Doppler Signatures Separation of FMCW Radar

Nowadays, Deep Convolutional Neural Networks (DCNNs) along with their outstanding advantages, are being widely used to identify human activities based on micro-Doppler (m-D) signatures obtained from radar sensors. In this study, we propose a new approach to improve the classification accuracy of living human activities by fully leveraging complicated m- D signatures corresponding to limb movements. The proposed ECM-Th-STFT separation algorithm is based on the energy concentration measure (ECM) and a Th-STFT filter with a threshold value of 0.3 to separate the m-D signal of the limbs from the Doppler signal related to the torso. Then, six DCNNs with structures ranging from simple to complex are utilized to extract features, recognize activities, and evaluate the separation efficiency of the proposed method. The experimental results show that the proposed method has improved classification accuracy by up to 6% compared to the original unseparated dataset.